Communicating a Neural Network Feature Vector (nnfv) to a Host and Receiving Back a Set of Weight Values for a Neural Network

1. A method comprising: (a) receiving a plurality of packets of a first flow onto a first device and generating therefrom a Neural Network Feature Vector (NNFV), wherein the first device includes a neural network, wherein the neural network includes multiple perceptron circuits, wherein the NNFV includes a log of information about the first flow, wherein the log of information includes a single-bit answer value, wherein the single-bit answer value is determined on the first device by a heuristic, and wherein the single-bit answer value is indicative of whether the first device has determined that the first flow has a particular type-of-flow characteristic; (b) communicating the NNFV from the first device to a second device, wherein the second device comprises a multi-layer software-implemented neural network, wherein the multi-layer software-implemented neural network comprises a first software perceptron and a second software perceptron, wherein the first software perceptron is in a first layer of the multi-layer software-implemented neural network, and wherein the second software perceptron is in a second layer of the multi-layer software-implemented neural network; (c) using the NNFV on the second device to determine a set of weight values for the neural network on the first device, wherein the using of (c) includes determining a first weight value of the set of weight values by determining a difference between the single-bit answer value and a first perceptron result value output by the first software perceptron, and wherein the using of (c) further includes determining a second weight value of the set of weight values by determining a difference between the single-bit answer value and a second perceptron result value output by the second software perceptron; (d) communicating the set of weight values from the second device to the first device; (e) loading the set of weight values into the neural network on the first device; (f) receiving a plurality of packets of a second flow onto the first device; and (g) using the neural network to make a determination on the first device that the second flow likely has the particular type-of-flow characteristic.

2. The method of claim 1, wherein the first device is a network flow processor, wherein the second device is a host processor that is coupled to the network flow processor by a bus, wherein the NNFV is communicated in (b) from the network flow processor across the bus to the host processor, and wherein the weight values are communicated in (d) from the host processor across the bus to the network flow processor.

3. The method of claim 2, wherein the host processor determines the set of weight values using the NNFV.

4. The method of claim 1, wherein prior to the loading of the set of weight values of (e) the first device had not received any packet of the second flow.

5. The method of claim 1, wherein the determination of the single-bit answer value by the heuristic in (a) involves consideration of a sum of packet sizes of packets of the flow during a predetermined period of time.

6. A method comprising: (a) receiving a plurality of packets of a first flow onto a first device; (b) from the plurality of packets of the first flow using a heuristic on the first device to generate a single-bit answer value, wherein the single-bit answer value is indicative of whether the first device has used the heuristic to determine that the first flow has a particular type-of-flow characteristic; (c) from the plurality of packets of the first flow generating a Neural Network Feature Vector (NNFV), wherein the NNFV includes a log of information about the first flow, and wherein the single-bit answer value is a part of the log of information; (d) communicating the NNFV from the first device to a second device, wherein the second device comprises a multi-layer software-implemented neural network, wherein the multi-layer software-implemented neural network comprises a software perceptron; (e) using the NNFV on the second device to determine a set of weight values for a neural network circuit on the first device, wherein the neural network circuit on the first device includes multiple perceptron circuits, wherein the using of (e) includes determining a weight value of the set of weight values by determining a difference between the single-bit answer value and a perceptron result value output by the software perceptron; (f) communicating the set of weight values from the second device to the first device; (g) loading the set of weight values into the neural network circuit on the first device; (h) receiving a plurality of packets of a second flow onto the first device; (i) from the plurality of packets of the second flow using the neural network circuit on the first device to determine a neural network determined value; (j) including the neutral network determined value into a flow entry for the second flow, wherein the flow entry is in a flow table; and (k) receiving another packet of the second flow onto the first device and using the flow entry in the flow table to carry out a predetermined action on said another packet.

7. The method of claim 6, wherein the single-bit answer value of (b) is indicative of whether the first flow is an elephant flow.

8. The method of claim 6, wherein the predetermined action of (k) is a routing of said another packet out of the first device.

9. The method of claim 6, wherein for each packet of the plurality of packets of the first flow the log of information includes a timestamp value.

10. The method of claim 9, wherein for each packet of the plurality of packets of the first flow the log of information includes a packet size value.

11. The method of claim 9, wherein the log of information includes a packet count value and a total size value.

12. The method of claim 6, wherein the first device is a network flow processor, wherein the second device is a host processor that is coupled to the network flow processor by a bus, wherein the NNFV is communicated in (d) from the network flow processor to the host processor across the bus, and wherein the set of weight values are communicated in (f) from the host processor to the network flow processor across the bus.

13. A method comprising: (a) receiving a plurality of packets of a first flow onto a first device and generating therefrom a Neural Network Feature Vector (NNFV), wherein the first device includes a neural network, wherein the neural network includes multiple perceptron circuits, wherein the NNFV includes a log of information about the first flow, wherein the log of information includes an answer value, wherein the answer value is determined on the first device by a heuristic, and wherein the answer value is indicative of whether the first device has determined that the first flow has a particular type-of-flow characteristic; (b) communicating the NNFV from the first device to a second device, wherein the second device comprises a multi-layer software-implemented neural network, wherein the multi-layer software-implemented neural network comprises a first software perceptron and a second software perceptron, wherein the first software perceptron is in a first layer of the multi-layer software-implemented neural network, and wherein the second software perceptron is in a second layer of the multi-layer software-implemented neural network; (c) using the NNFV on the second device to determine a set of weight values for the neural network on the first device, wherein the using of (c) includes determining a first weight value of the set of weight values by determining a difference between the answer value and a first perceptron result value output by the first software perceptron, and wherein the using of (c) further includes determining a second weight value of the set of weight values by determining a difference between the answer value and a second perceptron result value output by the second software perceptron; (d) communicating the set of weight values from the second device to the first device; (e) loading the set of weight values into the neural network on the first device; (f) receiving a plurality of packets of a second flow onto the first device; and (g) using the neural network to make a determination on the first device that the second flow likely has the particular type-of-flow characteristic.

14. The method of claim 13, wherein for each packet of the plurality of packets of the first flow the log of information includes a timestamp value and a packet size value, and wherein the first device is an integrated circuit.

15. The method of claim 13, wherein for each packet of the plurality of packets of the first flow the log of information includes a timestamp value and a packet size value, and wherein the NNFV is communicated in (b) across a network connection, and wherein the set of weight values is communicated in (d) across a network connection.